GNN classifier to distinguish between 𝜸 rays and CR.

ALPAQUITA-AS, the prototype of ALPACA, has been operating in situ since April 2023. It consists of 97 plastic scintillators, representing a quarter of the full ALPACA surface detector array. High-energy $\gamma$-ray astronomy, which involves measuring air shower particles at ground level, faces the challenge of discriminating $\gamma$ rays from the dominant background of cosmic rays (CR). In this work, we propose a classifier based on Graph Neural Network (GNN) to distinguish between $\gamma$ rays and CR. Following the success of Convolutional Neural Networks (CNN) in many areas and the deep learning revolution it spurred, interest in extending the use of convolution layers operating on graphs has grown. In some applications, GNNs have a notable advantage over CNNs because in certain scenarios the information is represented with graph structures naturally or more effectively. The training and evaluation datasets were generated through Monte Carlo simulations, the results are reported.